- Email: [email protected]
Analogs of thyrotropin-releasing hormone using an aminoglycine-based template
Peptides, Vol. 19, No. 10, 1679 –1683, 1998 Copyright © 1998 Elsevier Science Inc. Printed in the USA. All rights reserved 0196-9781/98 $19.00 1 .00
PII S0196-9781(98)00124-7
BRIEF COMMUNICATION
Analogs of Thyrotropin-releasing Hormone Using an Aminoglycine-based Template DEAN A. KIRBY,* WEI WANG,† MARVIN C. GERSHENGORN† AND JEAN E. RIVIER*1 *Clayton Foundation Laboratories for Peptide Biology, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, CA, 92037 †New York Hospital, Cornell University Medical Center, 525 E 68th Street, New York, NY 10021 Received 20 March 1998; Accepted 21 July 1998 KIRBY, D. A., W. WANG, M. C. GERSHENGORN AND J. E. RIVIER. Analogs of thyrotropin-releasing hormone using an aminoglycine-based template. PEPTIDES 19(10) 1679 –1683, 1998.—Novel thyrotropinreleasing hormone (TRH, pGlu-His-Pro-NH2) analogs, made by solid phase, were derived from the general scaffold pGlu-(D/L)Agl(X)-Pro-NH2 where Agl 5 aminoglycine. Analogs ranged from X being a proton to an acylating agent derived from substituted (aromatic heterocyclic rings) formic or acetic acids or an aminotriazolyl moiety (39-amino-1H-19,29,49-triazolyl) built on Na of aminoglycine or Nb of a,b-diaminoproprionic acid (Dpr). X was expected to mimic the electronic and structural characteristics of the imidazole ring of histidine. Analogs were purified by HPLC, characterized by mass spectrometry and isolated as either diastereoisomeric mixtures or pure isomers. Analogs, tested for their binding affinity to mouse pituitary TRH receptors, have apparent equilibrium inhibitory constants .1 mM. © 1998 Elsevier Science Inc. TRH analogs
TRH receptor
Betidamino acids
THYROTROPIN-RELEASING hormone (TRH), the first hypothalamic hypophysiotropic hormone to be characterized (1,2), controls the release of thyroid stimulating hormone (TSH) from the anterior pituitary gland, and is responsible for diverse pharmacological and physiological, central and peripheral actions [see review by O’Leary (11)]. Early reports originating from this laboratory described the synthesis of both TRH and analogs (15). In the present study we return to TRH, not so much to gain biologic insight, as to exploit this small bioactive molecule as a chemical template for exploring molecular diversity through the use of recently emerging methodologies. Orthogonally protected monoacylated aminoglycine derivatives have been identified as useful scaffolds for the introduction of chemical diversity into bioactive polypeptides (13). pGlu-(D/L)Agl(Fmoc)-Pro-resin was used as the template for the introduction of a number of acylating agents to yield TRH
Aminoglycine derivatives
analogs, which were screened for binding affinity at mouse pituitary TRH receptors (TRH-Rs). METHOD All tripeptides were synthesized manually using standard solid phase peptide synthesis techniques following Bocstrategy on methylbenzhydrylamine (MBHA) resins by methods previously described (7). pGlu-OH and BocPro-OH were purchased from Bachem (Torrance, CA), whereas the a-(tert-butyloxycarbonyl)-Na’-amino-(fluorenylmethyloxycarbonyl)-glycine [Boc-D/L-Agl(Fmoc)] and the methylated homolog [Boc-D/L-MeAgl(Fmoc)] were prepared by a modified method of Qasmi et al. using a-(isopropylthio)-Na’-(fluorenylmethyloxycarbonyl) glycine (6,12). Construction of the pGlu-Agl(Fmoc)-Pro-MBHA resin template proceeded in a straightforward manner using a twofold excess of protected amino acids and N,N9-diisopropyl-
1 Requests for reprints should be addressed to J. E. Rivier, The Salk Institute for Biological Studies, The Clayton Foundation Laboratories for Peptide Biology, 10010 N. Torrey Pines Road, La Jolla, CA 92037. E-Mail: [email protected]
1679
1680
carbodiimide in dichloromethane for 2 h. Following deprotection of the Agl sidechain using 20% piperidine in N-methylpyrrolidone, acylation was generally accomplished within 2 h by addition of a three-fold excess of the N-hydroxysuccidimide ester of the corresponding acid. Acylating agents were bought from Aldrich. Most compounds were rapidly and simultaneously produced by simply dividing the pGlu-Agl(NH2)-Pro-MBHA template, and acylating each aliquot individually with the appropriate activated carboxylic acid. Additionally, the orthogonally protected Agl scaffold was used to form an aminotriazole (Atz) by first treating the freed amine with diphenyl cyanocarbonimidate (PCI) followed by treatment with anhydrous hydrazine (18). Following Agl functionalization, the peptide amides were removed from the resin by treatment of anhydrous hydrofluoric acid for 1.5 h at 0°C, washed with ethyl ether, extracted in H20/20% CH3CN, and lyophilized. Crude peptide analogs were purified using a two-step preparative reverse-phase HPLC protocol using a highly resolutive linear triethylammonium phosphate/30% MeCN (pH 2.25) buffer system gradient followed by desalting in 0.1% TFA. Purified peptides were subjected to analysis using HPLC, capillary zone electrophoresis (CZE) and liquid secondary ion mass spectrometry (LSI-MS). As a consequence of using racemic Boc-Agl(Fmoc) or Boc-(Me)Agl(Fmoc) in the preparation of these analogs, it was not always possible to isolate each individual diastereomer; in those cases an approximate 50/50 mixture (seen analytically using CZE) of the two diastereomers was tested for binding affinity. Binding to AtT-20 mouse pituitary cells stably expressing TRH-Rs and in COS-1 cells transiently expressing TRH-Rs was used to screen for apparent “fit” between the receptors and the synthetic analogs. Competition binding to intact cells was performed for 3 h at room temperature with 2 nM [3H][N3-im-Me-His2]TRH and various concentrations of unlabelled TRH analogs. Cell monolayers were washed three times with cold buffer and binding was quantitated by first solubilizing cells with NaOH and counting in a liquid scintillation counter (8). RESULTS For the present study, eight novel structures were introduced on the pGlu-Agl(NH2)-Pro-NH2 template shown in Table 1. Also included for comparison were two related analogs (11 and 12) that employed pGlu-Dpr-Pro-NH2 as the template. Most analogs were obtained as powders and in expected yields following purification. In the simplest case (1), the sidechain of the Agl2 was left underivatized. In only two cases (7 and 8) were we able to isolate the individual diastereomers, so diastereomeric mixtures were generally tested for binding. Peptides were fully characterized by mass spectrometry. None of the 12 analogs reported here displaced 2 nM [3H][N3-im-Me-His2]TRH at a 1 micromolar concentration whereas TRH (1 mM) lowered binding of
KIRBY ET AL.
[3H][N3-im-Me-His2]TRH to less than 10% of total. Different retention times of analogs on HPLC reflect a broad range of lipophylicity values. DISCUSSION Early structure-activity studies of TRH by us and others indicated that the pyroglutamic acid and proline amide residues were essential for full hormonal (i.e., thyrotropin(TSH), and prolactin-releasing) activities, while slight modification of the histidine by similar aromatic functionalities was somewhat tolerated. Several analogs modified in the second position have been shown to display unusual high activity in the CNS, distinct from the hormonal activities (4,9,17). [N3-im-Me-His2]TRH and [Pyr(1)Ala2]TRH, both developed over twenty years ago, are still regarded as standards in many hormonal assays. It was concluded that the basicity and availability of p electrons imparted by the imidazole ring at position 2 were important for hormonal activity. More recently, groups have attempted to circumvent these electronic requirements using conformational constraints, producing several low affinity analogs (10). Aminoglycine (Agl)-based scaffolds have recently been identified as a convenient means for introducing chemical diversity into bioactive polypeptides through sidechain acylation using inexpensive and commercially available starting materials, producing what are referred to as “betides” (Scheme 1) (13). This study probes steric as well as basicity requirements for residue 2 of TRH. Most analogs were derived from the pGlu-Agl-Pro-MBHA scaffold, ranging in structure from an alanine-like or indole (trypophan-like) functionality (expected to be inactive) to a Pyr(1)Ala or imidazole (histidinelike) functionalities (expected to be active). Much to our surprise, no analog possessed significant binding affinity at concentrations up to 1 mM in an established assay (8). Because substitutions by basic amino acids such as Orn2, Lys2 and Arg2 had been shown to release TSH, although with low potency (15), we wondered whether [Agl2]TRH would bind to TRH-R. Any affinity would have suggested that Agl2 residue behaves like a basic amino acid; lack of affinity of [Agl2]TRH (1) would confirm the premise that Agl may be more Ala-like as suggested by earlier observations that Agl is a good alternative to Ala in a number of gonadotropin-releasing hormone antagonists (5) than lysine-like. [Ala2]TRH has been shown to be devoid of both TSH and central activity (13); likewise, [Agl2]TRH (1) was found to be inactive in our assay system. Acylation of both scaffolds (n 5 0 with N9H or N9Me) with the recently commercially available 4-imidazole-carboxylic acid yielded compounds 2 and 3 that are the closest homologues of TRH. We have two possible explanations for their lack of affinity in our binding assay. Electronically, we expect the pKa of the imidazole nitrogens to be modified by the additional conjugation with a carbonyl group. Steri-
ANALOGS OF TRH
1681 TABLE 1
CHEMICAL STRUCTURES OF TEMPLATE-BASED TRH ANALOGS WITH PHYSICAL AND BINDING PROPERTIES
a
Monoisotopic mass m/z, measured with LSI-MS; glycerol and 3-nitrobenzyl alcohol (1:1) matrix; Cs ion source. Retention times (min) determined by analytical HPLC using linear gradient conditions starting from 0% to 50% in 50 min; solvent system was A: 0.1% TFA; B: 25% CH3CN in 0.1% TFA. Column was Vydac C18 (5mM particle size; 2.1 3 150 mm); detection was 0.1 AUFS at 210 nm. Flow rate was 0.2 ml/min. c Binding affinities expressed as Ki values. d Two entries are given for both 7 and 8 representing the two resolved diastereomers. In all other cases, diastereomers coeluted or appeared as unresolved shoulders during their preparative purification. b
cally, we could show (Koerber, unpublished) that the volume spanned by the sidechain of residue 2 in analogs 2 and 3 is quite limited and further removed from the backbone as compared to the doughnut-shaped volume spanned by the imidazole ring of the native histidine sidechain. This is due to the planar nature of the amide bond that replaces the b-methylene group of histidine in TRH (13). We had hoped that extension by one methylene group as in compound 4 would have increased the chances for receptor binding. This was not the case. We therefore attempted to mimic the histidine sidechain with other readily available acylating agents. The substitution of a bulky indole sidechain, [Trp2]TRH, was shown to be inactive at doses up to 5000 times the effective dosage of TRH, probably due to the loss of basicity (16). To mimic this Trp substitution, we synthesized [Agl(indole-3-carboxylyl)2]TRH (5). Two other indolic analogs were prepared in the hope of detecting subtle conformational preferences by the TRH-R, resulting in the struc-
tural isomer [Agl(indole-2-carboxylyl)2]TRH (6) and the homolog [Agl(indole-3-acetyl)2]TRH (7). Speculating that the effects of electronegativity and heteroaromaticity in position 2 are important for hormonal activity, we then turned our attention toward modifying these parameters in the tripeptide template. Using straightforward acylation approaches produced [Agl(2-pyrazinecarboxylyl)2]TRH (8), [Dpr(2-pyrazine-carboxylyl)2]TRH (9), and [Agl(4-pyrazole-carboxylyl)2]TRH (10). None show affinity for the TRH receptor at 1 mM concentration. Another approach to an aromatic, nitrogen-containing sidechain similar to the imidazole ring of histidine with similar Lewis basicity, yielded 11 and 12 ([Agl/Dpr(3amino-1H-19,29,49-triazole)2]TRH) derived through two simple and high yield steps following the removal of the Fmoc sidechain protecting group as shown earlier (14). A similar modification on the sidechain of 4-aminophenylalanine at positions 5 and 6 of a gonadotropin-releasing hormone antagonist imparted a significant increase in duration
1682
KIRBY ET AL.
SCHEME 1. Preparation of betidamino acid-containing peptide using solid phase. In the example shown, b-Trp is formed by the acylation of an aminoglycine residue with indole-3-carboxylic acid (see Method for details).
of action (14). It is believed that the lack of affinity of these analogs for the TRH-R is the result of steric hindrance, basicity or hydrophobicity similar to that observed for [N1-im-Me-His2]TRH and [N3-im-Y-His2]TRH (Y 5 F, Cl, Br, I, CF3, CN or CO2H) (9) and unlike that observed for the very favorable [N3-im-Me-His2]TRH substitution (15). Finally, because all analogs were found to be inactive at the doses tested, the wide range of lipophilicity derived from their retention times on HPLC cannot be correlated with either potency or duration of action. In this series, 2 and 3 with a 3-imidazole-carboxylyl substitution are most resemblant to active TRH analogs, suggesting that the lack of binding of these analogs results from a stringent conformational requirement by pituitary TRH-Rs that shows extreme sensitivity to both size and basicity. Since these analogs contained an equal mixture of the L and D isomers, two areas of conformational space were simultaneously investigated (13). We are left to conclude that the presence of the additional amide linkage that tethers the backbone of the molecule to the sidechain functionality
must be disruptive to recognition and binding. Previously, the highly sensitive nature of receptor mediated hormonal activity was demonstrated by the stark difference in potency of isomeric pairs [N3-im-Me-His2]TRH (800% activity) versus [N1-im-Me-His2]TRH (0.04% activity) and [b-pyrazol1-yl-Ala2]TRH (150% activity) versus [b-pyrazol-3-ylAla2]TRH (5% activity) (3,15). In conclusion, we have demonstrated the value of aminoglycine-based scaffolds as they allowed for an economical exploration of a number of important parameters (steric and ionic) in the SAR of TRH. A variety of structures designed to mimic residue 2 of TRH (histidine) were prepared ranging from Ala-, His-, 3-pyridyl- and Trp-like yet, none showed significant affinity for TRH-R. ACKNOWLEDGEMENTS The authors thank C. Miller for HPLC and CZE analysis, Dr. A. G. Craig for mass spectral analysis, Dr. S. Koerber for computational analyses and D. Johns for manuscript preparation. This work was supported by NIH grants DK 26741 (JER) and DK 43036 (MCG) and the Hearst Foundation.
REFERENCES 1. Boler, J.; Enzmann, F.; Folkers, K.; Bowers, C. Y.; Schally, A. V. The identity of chemical and hormonal properties of the thyrotropin releasing hormone and pyroglutamyl-histidyl-proline amide. Biochem. Biophys. Res. Commun. 37:705–710; 1969. 2. Burgus, R.; Dunn, T.; Ward, D.; Vale, W.; Amoss, M.; Guillemin, R. De´rive´s polypeptidiques de synthe`se doue´s d’activite´ hypophysiotrope TRF. C. R. Acad. Sci. 268:2116 – 2118; 1969. 3. Coy, D. H.; Hirotsu, Y.; Redding, T. W.; Coy, E. J.; Schally, A. V. Synthesis and biological properties of the 2-L-b-(pyra-
zolyl-1)alanine analogs of luteinizing hormone-releasing hormone and thyrotropin-releasing hormone. J. Med. Chem. 18: 948 –949; 1975. 4. Faden, A. I.; Labroo, V. M.; Cohen, L. A. Imidazole-substituted analogues of TRH limit behavioral deficits after experimental brain trauma. J. Neurotrauma 10:101–108; 1993. 5. Jiang, G.; Miller, C.; Koerber, S. C.; Porter, J.; Craig, A. G.; Bhattacharjee, S.; Kraft, P.; Burris, T. P.; Campen, C. A.; Rivier, C. L.; Rivier, J. E. Betidamino Acid Scan of the GnRH Antagonist Acyline. J. Med. Chem. 40:3739 –3748; 1997. 6. Jiang, G.-C.; Simon, L.; Rivier, J. E. Orthogonally protected
ANALOGS OF TRH
7.
8.
9.
10.
11. 12. 13.
N-methyl-substituted a-aminoglycines. Prot. Pept. Lett. 3:219 –224; 1996. Kirby, D. A.; Boublik, J. H.; Rivier, J. E. Neuropeptide Y. Y1 and Y2 affinities of the complete series of analogues with single D-residue substitution. J. Med. Chem. 36:3802–3808; 1993. Laakkonen, L.; Li, W.; Perlman, J. H.; Guarnieri, F.; Osman, R.; Moeller, K. D.; Gershengorn, M. C. Restricted analogues provide evidence of a biologically active conformation of thyrotropin-releasing hormone. Mol. Pharmacol. 49:1092–1096; 1996. Labroo, V. M.; Vonhof, S.; Feuerstein, G. Z.; Cohen, L. A. Effect of histidine modification in TRH on binding to rat pituitary and brain receptors. In: Rivier, J. E., Marshall, G. R., Eds. Escom, Leiden, The Netherlands. 1989; pp. 127–128. Li, W.; Moeller, K. D. Conformationally restricted TRH analogs: The compatibility of a 6,5-bicyclic lactam-based mimetic with binding to TRH-R. J. Am. Chem. Soc. 118:10106 – 10112; 1996. O’Leary, R.; O’Connor, B. Thyrotropin-releasing hormone. J. Neurochem. 65:953–963; 1995. Qasmi, D.; Rene´, L.; Badet, B. An a-aminoglycine derivative suitable for solid phase peptide synthesis using Fmoc strategy. Tetrahedron Lett. 34:3861–3862; 1993. Rivier, J. E.; Jiang, G.-C.; Koerber, S. C.; Porter, J.; Craig,
1683
14.
15.
16.
17.
18.
A. G.; Hoeger, C. Betidamino acids: Versatile and constrained scaffolds for drug discovery. Proc. Natl. Acad. Sci. USA 93:2031–2036; 1996. Rivier, J.; Porter, J.; Hoeger, C.; Theobald, P.; Craig, A. G.; Dykert, J.; Corrigan, A.; Perrin, M.; Hook, W. A.; Siraganian, R. P.; Vale, W.; Rivier, C. Gonadotropin releasing hormone antagonists with Nv-triazolyl-ornithine, -lysine or -paraaminophenylalanine residues at positions 5 and 6. J. Med. Chem. 35:4270 – 4278; 1992. Rivier, J.; Vale, W.; Monahan, M.; Ling, N.; Burgus, R. Synthetic TRF (thyrotropin releasing factor) analogues 31,2: Effect of replacement or modification of histidine residue on biological activity. J. Med. Chem. 15:479 – 482; 1972. Sievertsson, H.; Chang, J.-K.; Folkers, K. On the role of the histidine moiety in the structure of the thyrotropin-releasing hormone. J. Med. Chem. 15:219 –221; 1972. Szirtes, T.; Kisfaludy, L.; Pa´losi, E.; Szporny, L. Synthesis of thyrotropin-releasing hormone analogues. 1. Complete dissociation of central nervous system effects from thyrotropinreleasing activity. J. Med. Chem. 27:741–745; 1984. Theobald, P.; Porter, J.; Hoeger, C.; Rivier, J. General method for incorporation of modified Nv-cyanoguanidino moieties on selected amino functions during solid-phase peptide synthesis. J. Am. Chem. Soc. 112:9624 –9626; 1990.
PII S0196-9781(98)00124-7
BRIEF COMMUNICATION
Analogs of Thyrotropin-releasing Hormone Using an Aminoglycine-based Template DEAN A. KIRBY,* WEI WANG,† MARVIN C. GERSHENGORN† AND JEAN E. RIVIER*1 *Clayton Foundation Laboratories for Peptide Biology, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, CA, 92037 †New York Hospital, Cornell University Medical Center, 525 E 68th Street, New York, NY 10021 Received 20 March 1998; Accepted 21 July 1998 KIRBY, D. A., W. WANG, M. C. GERSHENGORN AND J. E. RIVIER. Analogs of thyrotropin-releasing hormone using an aminoglycine-based template. PEPTIDES 19(10) 1679 –1683, 1998.—Novel thyrotropinreleasing hormone (TRH, pGlu-His-Pro-NH2) analogs, made by solid phase, were derived from the general scaffold pGlu-(D/L)Agl(X)-Pro-NH2 where Agl 5 aminoglycine. Analogs ranged from X being a proton to an acylating agent derived from substituted (aromatic heterocyclic rings) formic or acetic acids or an aminotriazolyl moiety (39-amino-1H-19,29,49-triazolyl) built on Na of aminoglycine or Nb of a,b-diaminoproprionic acid (Dpr). X was expected to mimic the electronic and structural characteristics of the imidazole ring of histidine. Analogs were purified by HPLC, characterized by mass spectrometry and isolated as either diastereoisomeric mixtures or pure isomers. Analogs, tested for their binding affinity to mouse pituitary TRH receptors, have apparent equilibrium inhibitory constants .1 mM. © 1998 Elsevier Science Inc. TRH analogs
TRH receptor
Betidamino acids
THYROTROPIN-RELEASING hormone (TRH), the first hypothalamic hypophysiotropic hormone to be characterized (1,2), controls the release of thyroid stimulating hormone (TSH) from the anterior pituitary gland, and is responsible for diverse pharmacological and physiological, central and peripheral actions [see review by O’Leary (11)]. Early reports originating from this laboratory described the synthesis of both TRH and analogs (15). In the present study we return to TRH, not so much to gain biologic insight, as to exploit this small bioactive molecule as a chemical template for exploring molecular diversity through the use of recently emerging methodologies. Orthogonally protected monoacylated aminoglycine derivatives have been identified as useful scaffolds for the introduction of chemical diversity into bioactive polypeptides (13). pGlu-(D/L)Agl(Fmoc)-Pro-resin was used as the template for the introduction of a number of acylating agents to yield TRH
Aminoglycine derivatives
analogs, which were screened for binding affinity at mouse pituitary TRH receptors (TRH-Rs). METHOD All tripeptides were synthesized manually using standard solid phase peptide synthesis techniques following Bocstrategy on methylbenzhydrylamine (MBHA) resins by methods previously described (7). pGlu-OH and BocPro-OH were purchased from Bachem (Torrance, CA), whereas the a-(tert-butyloxycarbonyl)-Na’-amino-(fluorenylmethyloxycarbonyl)-glycine [Boc-D/L-Agl(Fmoc)] and the methylated homolog [Boc-D/L-MeAgl(Fmoc)] were prepared by a modified method of Qasmi et al. using a-(isopropylthio)-Na’-(fluorenylmethyloxycarbonyl) glycine (6,12). Construction of the pGlu-Agl(Fmoc)-Pro-MBHA resin template proceeded in a straightforward manner using a twofold excess of protected amino acids and N,N9-diisopropyl-
1 Requests for reprints should be addressed to J. E. Rivier, The Salk Institute for Biological Studies, The Clayton Foundation Laboratories for Peptide Biology, 10010 N. Torrey Pines Road, La Jolla, CA 92037. E-Mail: [email protected]
1679
1680
carbodiimide in dichloromethane for 2 h. Following deprotection of the Agl sidechain using 20% piperidine in N-methylpyrrolidone, acylation was generally accomplished within 2 h by addition of a three-fold excess of the N-hydroxysuccidimide ester of the corresponding acid. Acylating agents were bought from Aldrich. Most compounds were rapidly and simultaneously produced by simply dividing the pGlu-Agl(NH2)-Pro-MBHA template, and acylating each aliquot individually with the appropriate activated carboxylic acid. Additionally, the orthogonally protected Agl scaffold was used to form an aminotriazole (Atz) by first treating the freed amine with diphenyl cyanocarbonimidate (PCI) followed by treatment with anhydrous hydrazine (18). Following Agl functionalization, the peptide amides were removed from the resin by treatment of anhydrous hydrofluoric acid for 1.5 h at 0°C, washed with ethyl ether, extracted in H20/20% CH3CN, and lyophilized. Crude peptide analogs were purified using a two-step preparative reverse-phase HPLC protocol using a highly resolutive linear triethylammonium phosphate/30% MeCN (pH 2.25) buffer system gradient followed by desalting in 0.1% TFA. Purified peptides were subjected to analysis using HPLC, capillary zone electrophoresis (CZE) and liquid secondary ion mass spectrometry (LSI-MS). As a consequence of using racemic Boc-Agl(Fmoc) or Boc-(Me)Agl(Fmoc) in the preparation of these analogs, it was not always possible to isolate each individual diastereomer; in those cases an approximate 50/50 mixture (seen analytically using CZE) of the two diastereomers was tested for binding affinity. Binding to AtT-20 mouse pituitary cells stably expressing TRH-Rs and in COS-1 cells transiently expressing TRH-Rs was used to screen for apparent “fit” between the receptors and the synthetic analogs. Competition binding to intact cells was performed for 3 h at room temperature with 2 nM [3H][N3-im-Me-His2]TRH and various concentrations of unlabelled TRH analogs. Cell monolayers were washed three times with cold buffer and binding was quantitated by first solubilizing cells with NaOH and counting in a liquid scintillation counter (8). RESULTS For the present study, eight novel structures were introduced on the pGlu-Agl(NH2)-Pro-NH2 template shown in Table 1. Also included for comparison were two related analogs (11 and 12) that employed pGlu-Dpr-Pro-NH2 as the template. Most analogs were obtained as powders and in expected yields following purification. In the simplest case (1), the sidechain of the Agl2 was left underivatized. In only two cases (7 and 8) were we able to isolate the individual diastereomers, so diastereomeric mixtures were generally tested for binding. Peptides were fully characterized by mass spectrometry. None of the 12 analogs reported here displaced 2 nM [3H][N3-im-Me-His2]TRH at a 1 micromolar concentration whereas TRH (1 mM) lowered binding of
KIRBY ET AL.
[3H][N3-im-Me-His2]TRH to less than 10% of total. Different retention times of analogs on HPLC reflect a broad range of lipophylicity values. DISCUSSION Early structure-activity studies of TRH by us and others indicated that the pyroglutamic acid and proline amide residues were essential for full hormonal (i.e., thyrotropin(TSH), and prolactin-releasing) activities, while slight modification of the histidine by similar aromatic functionalities was somewhat tolerated. Several analogs modified in the second position have been shown to display unusual high activity in the CNS, distinct from the hormonal activities (4,9,17). [N3-im-Me-His2]TRH and [Pyr(1)Ala2]TRH, both developed over twenty years ago, are still regarded as standards in many hormonal assays. It was concluded that the basicity and availability of p electrons imparted by the imidazole ring at position 2 were important for hormonal activity. More recently, groups have attempted to circumvent these electronic requirements using conformational constraints, producing several low affinity analogs (10). Aminoglycine (Agl)-based scaffolds have recently been identified as a convenient means for introducing chemical diversity into bioactive polypeptides through sidechain acylation using inexpensive and commercially available starting materials, producing what are referred to as “betides” (Scheme 1) (13). This study probes steric as well as basicity requirements for residue 2 of TRH. Most analogs were derived from the pGlu-Agl-Pro-MBHA scaffold, ranging in structure from an alanine-like or indole (trypophan-like) functionality (expected to be inactive) to a Pyr(1)Ala or imidazole (histidinelike) functionalities (expected to be active). Much to our surprise, no analog possessed significant binding affinity at concentrations up to 1 mM in an established assay (8). Because substitutions by basic amino acids such as Orn2, Lys2 and Arg2 had been shown to release TSH, although with low potency (15), we wondered whether [Agl2]TRH would bind to TRH-R. Any affinity would have suggested that Agl2 residue behaves like a basic amino acid; lack of affinity of [Agl2]TRH (1) would confirm the premise that Agl may be more Ala-like as suggested by earlier observations that Agl is a good alternative to Ala in a number of gonadotropin-releasing hormone antagonists (5) than lysine-like. [Ala2]TRH has been shown to be devoid of both TSH and central activity (13); likewise, [Agl2]TRH (1) was found to be inactive in our assay system. Acylation of both scaffolds (n 5 0 with N9H or N9Me) with the recently commercially available 4-imidazole-carboxylic acid yielded compounds 2 and 3 that are the closest homologues of TRH. We have two possible explanations for their lack of affinity in our binding assay. Electronically, we expect the pKa of the imidazole nitrogens to be modified by the additional conjugation with a carbonyl group. Steri-
ANALOGS OF TRH
1681 TABLE 1
CHEMICAL STRUCTURES OF TEMPLATE-BASED TRH ANALOGS WITH PHYSICAL AND BINDING PROPERTIES
a
Monoisotopic mass m/z, measured with LSI-MS; glycerol and 3-nitrobenzyl alcohol (1:1) matrix; Cs ion source. Retention times (min) determined by analytical HPLC using linear gradient conditions starting from 0% to 50% in 50 min; solvent system was A: 0.1% TFA; B: 25% CH3CN in 0.1% TFA. Column was Vydac C18 (5mM particle size; 2.1 3 150 mm); detection was 0.1 AUFS at 210 nm. Flow rate was 0.2 ml/min. c Binding affinities expressed as Ki values. d Two entries are given for both 7 and 8 representing the two resolved diastereomers. In all other cases, diastereomers coeluted or appeared as unresolved shoulders during their preparative purification. b
cally, we could show (Koerber, unpublished) that the volume spanned by the sidechain of residue 2 in analogs 2 and 3 is quite limited and further removed from the backbone as compared to the doughnut-shaped volume spanned by the imidazole ring of the native histidine sidechain. This is due to the planar nature of the amide bond that replaces the b-methylene group of histidine in TRH (13). We had hoped that extension by one methylene group as in compound 4 would have increased the chances for receptor binding. This was not the case. We therefore attempted to mimic the histidine sidechain with other readily available acylating agents. The substitution of a bulky indole sidechain, [Trp2]TRH, was shown to be inactive at doses up to 5000 times the effective dosage of TRH, probably due to the loss of basicity (16). To mimic this Trp substitution, we synthesized [Agl(indole-3-carboxylyl)2]TRH (5). Two other indolic analogs were prepared in the hope of detecting subtle conformational preferences by the TRH-R, resulting in the struc-
tural isomer [Agl(indole-2-carboxylyl)2]TRH (6) and the homolog [Agl(indole-3-acetyl)2]TRH (7). Speculating that the effects of electronegativity and heteroaromaticity in position 2 are important for hormonal activity, we then turned our attention toward modifying these parameters in the tripeptide template. Using straightforward acylation approaches produced [Agl(2-pyrazinecarboxylyl)2]TRH (8), [Dpr(2-pyrazine-carboxylyl)2]TRH (9), and [Agl(4-pyrazole-carboxylyl)2]TRH (10). None show affinity for the TRH receptor at 1 mM concentration. Another approach to an aromatic, nitrogen-containing sidechain similar to the imidazole ring of histidine with similar Lewis basicity, yielded 11 and 12 ([Agl/Dpr(3amino-1H-19,29,49-triazole)2]TRH) derived through two simple and high yield steps following the removal of the Fmoc sidechain protecting group as shown earlier (14). A similar modification on the sidechain of 4-aminophenylalanine at positions 5 and 6 of a gonadotropin-releasing hormone antagonist imparted a significant increase in duration
1682
KIRBY ET AL.
SCHEME 1. Preparation of betidamino acid-containing peptide using solid phase. In the example shown, b-Trp is formed by the acylation of an aminoglycine residue with indole-3-carboxylic acid (see Method for details).
of action (14). It is believed that the lack of affinity of these analogs for the TRH-R is the result of steric hindrance, basicity or hydrophobicity similar to that observed for [N1-im-Me-His2]TRH and [N3-im-Y-His2]TRH (Y 5 F, Cl, Br, I, CF3, CN or CO2H) (9) and unlike that observed for the very favorable [N3-im-Me-His2]TRH substitution (15). Finally, because all analogs were found to be inactive at the doses tested, the wide range of lipophilicity derived from their retention times on HPLC cannot be correlated with either potency or duration of action. In this series, 2 and 3 with a 3-imidazole-carboxylyl substitution are most resemblant to active TRH analogs, suggesting that the lack of binding of these analogs results from a stringent conformational requirement by pituitary TRH-Rs that shows extreme sensitivity to both size and basicity. Since these analogs contained an equal mixture of the L and D isomers, two areas of conformational space were simultaneously investigated (13). We are left to conclude that the presence of the additional amide linkage that tethers the backbone of the molecule to the sidechain functionality
must be disruptive to recognition and binding. Previously, the highly sensitive nature of receptor mediated hormonal activity was demonstrated by the stark difference in potency of isomeric pairs [N3-im-Me-His2]TRH (800% activity) versus [N1-im-Me-His2]TRH (0.04% activity) and [b-pyrazol1-yl-Ala2]TRH (150% activity) versus [b-pyrazol-3-ylAla2]TRH (5% activity) (3,15). In conclusion, we have demonstrated the value of aminoglycine-based scaffolds as they allowed for an economical exploration of a number of important parameters (steric and ionic) in the SAR of TRH. A variety of structures designed to mimic residue 2 of TRH (histidine) were prepared ranging from Ala-, His-, 3-pyridyl- and Trp-like yet, none showed significant affinity for TRH-R. ACKNOWLEDGEMENTS The authors thank C. Miller for HPLC and CZE analysis, Dr. A. G. Craig for mass spectral analysis, Dr. S. Koerber for computational analyses and D. Johns for manuscript preparation. This work was supported by NIH grants DK 26741 (JER) and DK 43036 (MCG) and the Hearst Foundation.
REFERENCES 1. Boler, J.; Enzmann, F.; Folkers, K.; Bowers, C. Y.; Schally, A. V. The identity of chemical and hormonal properties of the thyrotropin releasing hormone and pyroglutamyl-histidyl-proline amide. Biochem. Biophys. Res. Commun. 37:705–710; 1969. 2. Burgus, R.; Dunn, T.; Ward, D.; Vale, W.; Amoss, M.; Guillemin, R. De´rive´s polypeptidiques de synthe`se doue´s d’activite´ hypophysiotrope TRF. C. R. Acad. Sci. 268:2116 – 2118; 1969. 3. Coy, D. H.; Hirotsu, Y.; Redding, T. W.; Coy, E. J.; Schally, A. V. Synthesis and biological properties of the 2-L-b-(pyra-
zolyl-1)alanine analogs of luteinizing hormone-releasing hormone and thyrotropin-releasing hormone. J. Med. Chem. 18: 948 –949; 1975. 4. Faden, A. I.; Labroo, V. M.; Cohen, L. A. Imidazole-substituted analogues of TRH limit behavioral deficits after experimental brain trauma. J. Neurotrauma 10:101–108; 1993. 5. Jiang, G.; Miller, C.; Koerber, S. C.; Porter, J.; Craig, A. G.; Bhattacharjee, S.; Kraft, P.; Burris, T. P.; Campen, C. A.; Rivier, C. L.; Rivier, J. E. Betidamino Acid Scan of the GnRH Antagonist Acyline. J. Med. Chem. 40:3739 –3748; 1997. 6. Jiang, G.-C.; Simon, L.; Rivier, J. E. Orthogonally protected
ANALOGS OF TRH
7.
8.
9.
10.
11. 12. 13.
N-methyl-substituted a-aminoglycines. Prot. Pept. Lett. 3:219 –224; 1996. Kirby, D. A.; Boublik, J. H.; Rivier, J. E. Neuropeptide Y. Y1 and Y2 affinities of the complete series of analogues with single D-residue substitution. J. Med. Chem. 36:3802–3808; 1993. Laakkonen, L.; Li, W.; Perlman, J. H.; Guarnieri, F.; Osman, R.; Moeller, K. D.; Gershengorn, M. C. Restricted analogues provide evidence of a biologically active conformation of thyrotropin-releasing hormone. Mol. Pharmacol. 49:1092–1096; 1996. Labroo, V. M.; Vonhof, S.; Feuerstein, G. Z.; Cohen, L. A. Effect of histidine modification in TRH on binding to rat pituitary and brain receptors. In: Rivier, J. E., Marshall, G. R., Eds. Escom, Leiden, The Netherlands. 1989; pp. 127–128. Li, W.; Moeller, K. D. Conformationally restricted TRH analogs: The compatibility of a 6,5-bicyclic lactam-based mimetic with binding to TRH-R. J. Am. Chem. Soc. 118:10106 – 10112; 1996. O’Leary, R.; O’Connor, B. Thyrotropin-releasing hormone. J. Neurochem. 65:953–963; 1995. Qasmi, D.; Rene´, L.; Badet, B. An a-aminoglycine derivative suitable for solid phase peptide synthesis using Fmoc strategy. Tetrahedron Lett. 34:3861–3862; 1993. Rivier, J. E.; Jiang, G.-C.; Koerber, S. C.; Porter, J.; Craig,
1683
14.
15.
16.
17.
18.
A. G.; Hoeger, C. Betidamino acids: Versatile and constrained scaffolds for drug discovery. Proc. Natl. Acad. Sci. USA 93:2031–2036; 1996. Rivier, J.; Porter, J.; Hoeger, C.; Theobald, P.; Craig, A. G.; Dykert, J.; Corrigan, A.; Perrin, M.; Hook, W. A.; Siraganian, R. P.; Vale, W.; Rivier, C. Gonadotropin releasing hormone antagonists with Nv-triazolyl-ornithine, -lysine or -paraaminophenylalanine residues at positions 5 and 6. J. Med. Chem. 35:4270 – 4278; 1992. Rivier, J.; Vale, W.; Monahan, M.; Ling, N.; Burgus, R. Synthetic TRF (thyrotropin releasing factor) analogues 31,2: Effect of replacement or modification of histidine residue on biological activity. J. Med. Chem. 15:479 – 482; 1972. Sievertsson, H.; Chang, J.-K.; Folkers, K. On the role of the histidine moiety in the structure of the thyrotropin-releasing hormone. J. Med. Chem. 15:219 –221; 1972. Szirtes, T.; Kisfaludy, L.; Pa´losi, E.; Szporny, L. Synthesis of thyrotropin-releasing hormone analogues. 1. Complete dissociation of central nervous system effects from thyrotropinreleasing activity. J. Med. Chem. 27:741–745; 1984. Theobald, P.; Porter, J.; Hoeger, C.; Rivier, J. General method for incorporation of modified Nv-cyanoguanidino moieties on selected amino functions during solid-phase peptide synthesis. J. Am. Chem. Soc. 112:9624 –9626; 1990.